UK: Zdenek Kalal, a PhD student at the University of Surrey in the UK has developed a novel camera that he calls Predator. The Predator system can be used to follow objects; for example, it can track the tip of a pen that might be held in the mouth of a paralysed person, who could use it to control a computer mouse.
It is pertinent to mention here that Microsoft also has recently showcased OneVision Video Recognizer. But, in contrast with other tracking techniques, Predator learns from its mistakes and is then able to make better decisions in the future — resulting in a real-time visual tracking system that improves its performance over time.
In the video, he demonstrates it by selecting his finger as it is filmed by the camera. The selection tracks his hands and follows it almost perfectly. What makes it amazing is that it learns from the movement and the tracking improves. With something like this we could have truly “Minority Report” style human-computer interface.
Zdenek Kalal has been recently presented with the ‘Technology Everywhere’ award at a ceremony hosted by EPSRC, the UK’s research funding council, and IBM, for his work on a smart camera tracking techniques. His project entitled ‘Predator: A Smart Camera that Learns from Experience’.
Source: University of Surrey